For states that participated in the medium system portion of the survey, the data collection process was similar to
that of large systems. Survey instruments were provided to the states for the medium systems in each state’s
sample. States worked with the systems to complete the questionnaire and collect and prepare documentation for
the projects listed on the questionnaire.
Once the need for medium systems in the fully surveyed states was calculated, EPA used it to estimate the need
for the partial participation states. An average need per system was calculated by stratum using data from the fully
participating states and applied to the inventory of systems in the partial participation states. The inventory of
systems was verified by state personnel.
Of the 1,892 medium systems that were randomly selected and received a survey, responses were received for
1,884 systems for a response rate of 99.6 percent. Based on the survey response and migration of systems between
strata, 1,918 of the survey responses were for systems that serve 3,301 to 100,000 people. The number of medium
systems in the final sample for each state, Puerto Rico, the District of Columbia, and the U.S. territories is shown
in Exhibit A.3.
Assessing the Need for Small Systems
The infrastructure need reported for small systems serving 3,300 people or fewer is based on the findings of the
2007 Assessment when data were collected through a field survey of 600 small water systems. The 2007 small
system sample was designed to estimate the national need of small systems with a 95 percent confidence interval
of plus or minus 25 percent. Because the field survey data are believed to be highly accurate and due to resource
constraints, EPA did not survey these systems again in 2015. Instead, EPA used the projects reported for the 2007
Assessment, applied the 2015 cost models, converted all costs to 2015 dollars, calculated the average need per
small system, and multiplied this average by the number of small systems in the 2015 inventory to estimate the
2015 needs for these systems.
Assessing the Need of Not-for-Profit Noncommunity Water Systems
NPNCWSs are eligible for DWSRF funding. The 2015 estimate of need of NPNCWSs was based on the findings
of the 1999 Assessment, which surveyed a statistical sample of these systems. These findings were adjusted to
January 2015 dollars using the Construction Cost Index. The national need for NPNCWSs was allocated among
the states in proportion to the 1999 inventory of NPNCWSs in each state in a manner similar to that used for small
systems with an adjustment to 2015 dollars.
During the 1999 Assessment, EPA collected data from a national sample of 100 NPNCWSs through site visits.
Unlike the sampling design for CWSs, the NPNCWS sample was not stratified into size and source categories
because EPA lacked the empirical information on variance necessary for developing strata. The sample used for
the 1999 Assessment for NPNCWSs was designed to provide a 95 percent confidence interval that is within a
range of ± 30 percent of the estimated need.
Very little information about the needs of NPNCWS’s was available before the 1999 Assessment, including data
on the number of not-for-profit systems, the type of infrastructure needed, or their total need. The cost of collecting
this information is high because there are a large number of non-community systems and site visits would be
required to collect information about their projects and needs. Despite their large number, it was expected that
their total need would be relatively low and would have a small impact on the estimate of the national and each
state’s needs. In 1999, EPA selected a small sample of NPNCWSs to minimize the cost of collecting these data.
Therefore, the margin of error for NPNCWSs is larger than for CWSs. In addition to the high margin of error, the